The predominant view among philosophers of language is that proper names are directly referential and contribute only the referent itself to the proposition expressed by sentences in which they occur. It has, however, long been recognized that certain names, or uses of names, resist this kind of analysis, and various alternative, descriptivist frameworks have been developed that seem to provide solutions to many of the problems associated with direct reference views (while attempting to accommodate the evidence in favor them). Here, I focus primarily on two-dimensional solutions. I argue that although two-dimensional systems produce elegant and attractive results, these results – or at least strikingly parallel results – can, in fact, be achieved without taking on board the descriptivist assumptions adopted or special tricks developed by two-dimensionalists. To achieve the desired results, however, we need to make certain other modifications to the standard referentialist picture – including recognizing that proper names can make different semantic contributions in different situations – and I explore what independent motivations we have for adopting those modifications, as well as advantages (many) and costs (ultimately low) of making them.